Axoniq, which started out 15 years ago making event-sourcing architecture, has made a hard launch of the first unified platform with a native explainability layer. The Axoniq Platform gives enterprises in complex and regulated industries the ability to integrate AI and distributed systems into their legacy infrastructure in a controlled, transparent way.
“The company started up about 15 years ago out of a consulting firm, doing event-sourcing architecture, which concerns the production and detection of events, and that by design, it collects system decisions in way of command, query and event, which means the decision it has made natively,” said Jessica Reeves, CEO of AxonIQ. “So every single decision, whether it’s a tiny, little one or a huge one, is stored and then it flows into what we call an event store, and then you kind of forget everything else by design.” Back then, our co-founder, Allard [Buijze], he started it 15 years ago out of a consulting firm called Trifork, and he realized, wow, this is super powerful. At that point in time, AI wasn’t really prominent for distributed systems.”
Reeves noted that the emphasis with event-driven architecture was on explainability and replayability, and that its performance and scalability made its strength primarily with complex and dynamic workloads, with regulated industries and e-commerce standing out as key use cases.
“It’s what happened versus why it happened, what sequence that happened, how it happened, who made that change, and so forth,” Reeves said. “So we have a very deep lineage — kind of the forensics of the explainability.”
The Axoniq Platform preserves every business decision with a complete causal history, providing full explainability with context, memory, and accountability. It closes the trust gap that has kept many organizations from successfully embracing AI.
“Trust is the bottleneck in enterprise AI,” Reeves said. “Without memory, explainability, and observability, AI becomes a risk. We built Axoniq Platform so teams can deploy AI-enabled systems that are traceable, testable, and aligned with business and compliance needs. As organizations race to integrate AI, our new platform marks a fundamental shift in how they approach the task. We turn legacy infrastructure into AI-ready systems where AI can show their work and have meaningful context fed to them to make better decisions.”
Axoniq redefines AI explainability as a systems architecture challenge rather than a model-level feature. Traditional approaches attempt to explain model outputs in isolation, missing the wider decision context across services, workflows, and human inputs. Axoniq captures the complete causal chain of events that lead to an AI decision, before, during, and after model inference, enabling continuous traceability, auditability, and governance. This systems-first approach overcomes the limitations of post–hoc explainability tools and meets the real demands of regulators asking, “Why did this decision happen?”
The actual company that became AxonIQ started about eight years ago, Reeves noted.
The AxonIQ platform has three primary components. One is open source Axon Framework 5, which is really its long standing base. The second is Axon Server 2025.2.0, although you are not limited to the use of the Axon Server if yould prefer another one. The third component is agents.
“The components were all kind of everywhere, and now I would say they’re all consolidated into a singular platform. You can see the platform kind of walks through the whole journey of how to develop, of the whole software life cycle. So the underbelly of the platform is the framework. When it comes to how do the components work together? That’s the framework powering this platform. You can simply get started and you know what you’re doing. You can just download the framework and get going. You can learn about it. But the biggest thing that we realized was event sourcing. Architecture is a complex topic, and not many, even mid-level type folks understand it. So we wanted to help people start building right away, without having to be an expert in event sourcing architecture. These are some of my previous projects, but you can start a new project. You can see it starts with different templates, if you will. But you also can just create a sales development tracker application. And then you can see here, from feeding into the framework itself, it then generates this journey, and it walks you step by step. If you are not technical, you could vibe code. If you are more technical, you can figure out the requirements here. So we do let you kind of pick your path.”
The agents enter the process at this stage.
“You also have a development agent here that, and this is where the agents come in,” Reeves said. So it’s all intertwined, and the agents then help you, kind of hold your hand through that journey, of like, this is what I’m doing. This is what that means. You can see that as you’re thinking through what we call journeys, which is really the user flow of what happens, and you can see, okay, you have to create a lead view, leads, update leads, etc, just based off of my prompt. You can go in look at it. It’s all colour coded, so you can actually see what’s happening, and kind of learn the logic of it. And you can approve and deny these. You also can suggest a new alternative flow. You can see this goes through the schemas, the event handlers, etc, so it really shows you the logic here again. You can see related journeys. You can approve the development agent. You can ask questions like, ‘why is this happening’? it then spits out all of the code, including test files and API files. I chose Kotlin as my language. Upcoming, it will also be Java, Kotlin, and in the roadmap will be other languages as well. So we’ll start leaning more into language agnostic. And so that’s where kind of the agent side comes in, as well as analyze. So it’s all congruent to the journey. So if you build an application, in my example, sales development agent or application, you can then analyze it. There’s an agent right beside it that says, What is going on. Why did this happen? And because the nature of event sourcing, which is getting all of that plumbing from the framework itself, it can answer the questions in a really powerful way. It follows you through the platform, from learning to then building applications. If you already have something built and it’s brownfield code, you can then, in the roadmap we’re coming up with, plop your kind of legacy code into this builder, and it will help you refactor it as well. That framework powers the whole platform. You have the option to use our server as an event store or someone else’s servers, and we have different API packets for that. You can configure that as well in the platform. And then the agentic and insights layer comes with the work you’re actually doing with the framework and the server piece.”
“Event-sourcing has always been the backbone of Axon Framework. It gives systems memory, a trustworthy history, and the ability to evolve as business needs change,” said Allard Buijze, Founder and CTO of Axoniq. “Fourteen years later, it’s gratifying to see it evolve into the ideal backend for AI, which needs context, history, and causality. Axoniq delivers exactly that.”
Reeves said that they did a soft launch during their conference October 1, and then we did a hard launch on Halloween. “So it is out, it’s ready,” she said. “Obviously, we’ll iterate. We’ll make it better based off feedback. So there’s a feedback button to what do people want, what do they like, what they don’t like. So we’ll iterate along the way, throughout and obviously, given that it is a platform, we can push out those iterations quite quickly. We’re also working on different components, where a lot of folks that that are customers and users come from regulated industries, so there’s even some of the components are available on-prem. Our kind of end goal is that pick your poison. Do you want it all cloud based on your cloud? Or you can download it in a way that’s on-prem to you as well.”
Four core capabilities define the Axoniq Platform:
Persistent Event-Based Memory provides full visibility into the decision context of every AI action, enabling time-travel debugging and compliance-ready traceability.
Dynamic Consistency Boundary (DCB) allows organizations to redefine data and transactional boundaries without losing historical event data—cutting data evolution timelines from quarters to weeks.
AI Observability delivers real-time tracing and explainability built into the event stream itself, ensuring every business decision is captured and auditable.
Agent-Ready Runtime provides enterprise-grade infrastructure to safely power autonomous systems with Model Context Protocol (MCP) extensions that let AI agents interact securely with enterprise systems.
Reeves said that DCB is likely the most interesting thing here.
“Basically, DCB enables you to edit the architecture in a meaningful way, without having to rebuild everything that was built on top of that architecture. Developers don’t have to do all that work again. The business perspective also doesn’t have to invest in all that work again. So that’s it’s really exciting, the DCB aspect of it, and it’s a rather new concept in the last year or so for us. And again, I would say AI is the catalyst of like, why now it’s all AI. It’s the catalyst of all of us. It’s changing all of our worlds today. We say it’s not a wave, but it’s like a tsunami that’s coming at us. And are we ready for it? I would argue most companies aren’t. So that is absolutely the reason why all of this is coming to the forefront.”
From a framework perspective, Reeves said that the company has had about 73 million downloads of the framework itself in the last two years.
“It’s a pretty large footprint in terms of contributors,” she indicated. “It’s a tight knit community. I would say when you’re comparing the open source contributor communities, it’s probably a little bit more niche, but when you look at the actual usage globally, it’s huge. As I said, it’s 73 million downloads in the last two years. That represents about 65,000 companies, 80% of the Fortune 100 and about 60% of the Fortune 500. I think again, AI is the catalyst. I think we’re going to see it more and more. We are seeing a lot in like regulated industries that really, really need that log or archive of what’s going on so financial, insurance, health, government, but also highly complex systems like E commerce, of fulfillment, pricing, shipping, all of that. Those are kind of the main industries that we see the patterns in from an open source perspective.”
Reeves said that Axoniq will continue to stand with those open source partners.
“Our open-source foundation is what makes this possible,” Reeves stated. “We’re not moving away from the community that built Axoniq. We’re building the future with them.”
So how does AxonIQ get this to market and sell it most effectively?
“For us, we think of, it’s all about value,” Reeves stated. “Part of bringing it to market was really the platform to help lower the barrier of entry. So you saw with the agent, I interact with it and built a code that was event driven by nature. I’m not a technologist, whereas we think that that’s super powerful for people that don’t need to know event driven architecture to kind of give it to the masses in that way. So that’s one strategy, to widen the aperture in terms of the persona that we’re going after. I also think when you think about different partnerships we have under NDA, that we’re cooking up a few partnerships, I have a very adjacent view when you think about, there’s a lot of event streaming type of folks out there. Event sourcing is different. There is some overlap, because we do stream as well, but thinking about from a go to market perspective, how do we give them the power of event sourcing as well? Partnerships are a huge play in that. In addition, while holding their hand is huge, so is just getting the word out there. I think that historically, we’ve been very focused on the technical, the very, very technical, nitty gritty of it. I think amplifying the business value and benefits of it will help a lot and into those personas that care about it, like the CTO, the CIO, even the CFO, CIO, will care about the efficiency and gain that you get. We have a large Swedish furniture retail company that got 2,400% performance gains by switching over to us. So to me, that catches the eye of as an operator, who will say ‘I need more of that.’”
This is also a big deal for the channel, Reeves said.
“For me, it’s very unique in that for the channel, either you’re going to tell your users and customers to basically do it yourself using 10 different technologies that you then have to stitch together and maintain, or there’s someone that’s been doing it for well over a decade and a half that’s experts in this, and you can go into one seamless place and do that. So to me, it’s having one kind of channel partner versus having to figure out a handful of different players to come along. Because I do think this use case is becoming more and more powerful today.”
Axoniq Platform, which includes Axon Framework 5, Axon Server 2025.2.0, and Agents, is available at www.axoniq.io with full documentation, tutorials, and sample applications.